Local dockerfile for hassle free reproducible paper

I just published this sketch2svg notebook for two reasons.
First, the algorithm proposed by these researchers gives great results.
Second, this illustrates how deepnote can ease the way towards reproducible science.

Indeed, it is a fact that it is more and more common or even expected to give access to code, data, and set up details when an article is published. This is great. But there is still a lot of friction to quickly get to the experiment part because it is not simple to get everything working together.

The image processing online review does a great job to allow quick experiments with an input/output interaction (and also parameter tuning). See for example this demo. These papers are definitely reproducible, however the execution environment is abstracted away from the user, so that it limits the ability to experiment further (with batches of experiments for example).

To make the environment closer to the user, docker comes handy. But not everyone is (or should be) well versed in compilation and docker environments.
With the local dockerfile mechanism in deepnote, it is possible for one person to do the setup and compilation so that everyone can benefit and learn from it (if needed). Good job deepnote!
Like Comment

The san pellegrino label moiré effect

Hey, check this notebook to reverse engineer the moiré effect present in the San Pellegrino water label.
(this is also an entry for deepnote publishing competition)
Like Comment